SEO in the Age of AI and Long-Form Search
Google’s Danny Sullivan and John Mueller recently offered guidance on ranking in LLM-based search and chat via the Search Off The Record podcast. One widely circulated piece of advice—“chunk your content”—was debunked, highlighting that this isn’t the approach SEOs should focus on.
The Next Generation of Search
Historically, Google ranked content based on keywords, with PageRank leveraging anchor text to gauge authority. The 2012 introduction of the Knowledge Graph marked a shift from ranking by strings to ranking by entities—moving toward understanding the real world.
Today, Google describes this as the next generation of search, tapping into collective web intelligence and understanding queries more like humans do.
While the underlying infrastructure remains Google Search, answers now appear in long-form formats, often addressing multiple follow-up questions beyond the initial query. The old paradigm of optimizing for a single keyword per page is fractured by query fan-out—multiple queries may now surface the same content differently.
Writing for Long-Form AI Answers
Some SEOs suggest breaking pages into “bite-sized chunks,” believing AI reads content more effectively this way. Google engineers caution against this:
-
Well-structured pages with headings, lists, and HTML formatting are already logically “chunked” for humans and machines.
-
Crafting multiple versions of content—one for humans, one for LLMs—is unnecessary.
Danny Sullivan said:
“We really don’t want people to be crafting anything for Search specifically… We want you to craft content for humans.”
The takeaway: write for humans first, not for AI systems. Over-optimizing for machines is a short-term tactic; systems evolve to reward human-focused content, making machine-specific tweaks often obsolete.
The Referral Challenge and Query Fan-Out
One under-discussed impact of AI search is query fan-out, which affects referral traffic. A single user query can now trigger multiple AI-generated answers, but the websites selected often don’t measure up in expertise or coverage.
This can reduce referral opportunities for subject-matter experts, making visibility in traditional search more fragmented.
Focus on the Big Picture
Danny emphasizes that small, machine-specific optimizations rarely deliver long-term results. Instead, SEOs should focus on foundational goals:
-
Produce clear, accurate, and human-readable content
-
Build authority and relevance through comprehensive coverage
-
Prioritize expertise, trustworthiness, and user satisfaction
The Problem of Hidden Expert Content
A growing concern is that expert publications are increasingly buried in Google’s results, often hidden behind the More tab or requiring clicks through the News tab. This reduces visibility for authoritative sources and can lead to arbitrary or low-quality AI-generated search results dominating SERPs.
The implication is clear: ranking in AI-driven search is not about manipulating systems—it’s about maintaining expertise, authority, and human-first content quality, while understanding how search visibility is shifting in a multi-query, AI-supplemented landscape.
Google’s AI Mode Surfaces Low-Quality Content
This example wasn’t cherry-picked—it came from a straightforward query about styling a sweatshirt. The results in Google’s AI Mode highlight a worrying trend: sites lacking expertise or authority dominate the answer set.
The AI cited the following pages:
-
Abandoned Medium Blog (2018) – Only two posts, both with broken images. Not authoritative.
-
LinkedIn article – A social networking platform, not a trusted source for expert style advice.
-
Sneaker retailer blog – Commercial site, not focused on fashion expertise.
These results show that Google’s AI Mode can prioritize visibility over credibility, surfacing content that fails to meet expert or authoritative standards.
Google Hides Expert Content Behind the “More > News” Tab
In many cases, Google’s AI search does not surface high-quality, authoritative sources by default. For example, queries that could have highlighted expert content from GQ or The New York Times instead show lower-quality pages.
To access reputable content, users often have to click More > News, effectively burying expert sources and reducing their visibility.
This behavior highlights a growing challenge: expert content is being deprioritized, while AI-generated or less authoritative pages take prominence in the default view.
GEO or SEO—It Doesn’t Matter
Debates about GEO, AEO, or whether “it’s all SEO” are largely hand-waving. The real issue is that high-quality sites are being buried, losing traffic, and struggling to survive.
Low-quality search results dominate the SERPs, leaving little joy in discovery. When was the last time you stumbled upon a genuinely valuable site you wanted to share? Instead, Google often surfaces garbage stacked on more garbage.
It’s time for a reset. Google could prioritize original, high-quality search results for users, while experimental AI and Gemini-style features live under a “More” tab.
[Listen to the podcast here]
